Mortgage recapture rate, benchmarked.
A straight answer on what a good mortgage recapture rate is, how Rocket's number compares to the industry baseline, how to calculate your own, and the four levers that actually move the ratio.
The bands you are grading yourself against.
Depository and IMB average. The number cited most often by vendors and analysts.
A loan officer running a disciplined past-client workflow without an intelligence tool.
Shops with servicing retention, structured retention workflow, and an intelligence layer.
Rocket Mortgage's publicly disclosed band on servicing-retained books. The ceiling to aim at.
Bands are rolling twelve-month averages. Volatility within a quarter can be large, especially in a flat or rising rate environment.
How to calculate your recapture rate.
(past borrowers whose next mortgage was with YOU)
/ (past borrowers whose next mortgage was with anyone)
over a rolling twelve-month window
The denominator is the hard part. Three ways loan officers measure it:
- Deed-of-trust records. Public records show a new mortgage recorded against a property. Reliable but lags by weeks to months and requires a county-by-county data layer.
- Servicing payoff data. If the lender services the loan, a payoff request and incoming wire identify when (and sometimes to whom) the borrower refinanced. Reliable and fast, but only works for serviced loans.
- Third-party enrichment. Vendors like ATTOM, DataTree, and CoreLogic maintain nationwide databases of recorded deeds. Costs money but closes the measurement gap for non-serviced books.
LO Radar handles the measurement layer automatically for the covered county footprint and surfaces the attribution dashboard on every Performance and Enterprise account.
What actually moves a recapture rate.
1. Daily monitoring, not monthly scans.
ICE data cited through 2026 shows that 78 percent of borrowers shop one or two lenders before closing. Whichever loan officer reaches the borrower first typically wins the call. A monthly manual scan of the servicing report cannot keep up; daily automated monitoring can. Rate Trip Wires in LO Radar pre-stage outreach for rates the borrower is not yet in the money on and release the batch the week the trigger prints, so the LO is first by design.
2. Twelve-month ARM lead time, not three.
At T minus three months, the client is in panic mode and shopping. At T minus twelve, nobody else is calling. Starting the ARM-reset conversation at T-12 turns into a 3x conversion improvement on the ARM subbook. Every 5/1, 7/1, and 10/1 ARM in your book has a known reset date; the LO Radar ARM Reset Calendar runs the T-12 / T-9 / T-6 / T-3 / T-1 cadence automatically.
3. Non-rate triggers so you are still calling in a flat market.
A pure rate-drop recapture workflow dies in a 7 percent market. The borrowers you need to call now are the ones with equity, ARM resets, HELOC headroom, PMI nearing cancellation, and anniversary windows. LO Radar runs seven trigger categories plus four newer capabilities (Rate Trip Wires, Permit Radar, Program Eligibility, Outreach Monitor) so the daily call list stays non-empty regardless of the rate environment.
4. Pre-drafted outreach in the LO's own voice.
The delta between identifying the opportunity and writing the message is where recapture rate dies in practice. If every rate-drop surfaced requires thirty minutes of writing, most never go out. If every opportunity ships with a draft in the LO's own voice and RESPA / TILA / TCPA posture inline, the LO ships. Voice Training in LO Radar learns the LO's tone from fifteen to thirty prior messages and renders every draft in three variants.
Measure your recapture rate against the bands above.
Bring a CSV from your LOS; LO Radar computes the baseline and shows the top three calls that would move the ratio in the first week.
Book the 15-minute walkthrough